Examination of the structure-activity relationships (SARs) of several of the classic stimulants provides not only an understanding of the development of other drugs, but provides important clues as to the underlying mechanisms involved in interaction with the target protein(s). The following sections will hopefully illustrate both of these points.
There are a number of related structures that are often referred to as "amphetamines," although the name amphetamine refers to one specific molecular entity. Grouped in this class would be (+)-amphetamine (10), iV-methyl-
amphetamine [S-(+)-methamphetamine, 18], phentermine (19), phenmetrazine (Preludin, 20), and phendimetrazine (21). Diethylpro-pion (Tenuate; 16) is used as an appetite suppressant and, although it has the amphetamine skeleton, its effects are much weaker as a stimulant than those of the other structures listed here. The stereochemistry at the a side-chain methyl group is the same for the most potent enantiomer of each structure, although the pure enantiomer has not generally been marketed except for the cases of (+)-amphet-amine (10)and (+)-methamphetamine (18).
The structural requirements of the dopamine (and norepinephrine) transporter appear to be fairly rigid. There is very little molecular variation that is tolerated without significant loss of activity. The relatively limited information that is available, mostly from animal studies, can be summarized by considering the various areas of substitution for a general phenethylamine structure. These structure-activity relationships have been recently surveyed (132), and an extensive and comprehensive review by Biel and Bopp (133) covered the older literature.
5.1.1 Length of the Side-Chain. The length of the side-chain is limited to two carbon atoms (134, 135). That is, for transporter substrates, the optimum pharmacophoric template appears to be a basic nitrogen two carbon atoms removed from an aromatic ring system. This observation of course is not too surprising, given that the transporter substrates do-
norepinephrine, and serotonin all bear this essential core.
5.1.2 Nitrogen Substituents. Nitrogen sub-stituents are very limited. The primary amine (amphetamine)and the AT-methylamine (meth-amphetamine) are the most potent compounds (135). An N-methyl increases the potency of both amphetamine and cathinone (2) (136). Larger alkyl groups (135,137) or N,N-dialkylation, either dramatically attenuate or completely abolish stimulant activity (138). Nevertheless, iV,iV-dimethylamphetamine has appeared on the illicit market (139) and does appear to have behavioral effects in rats and monkeys similar to amphetamine (138, 140). The rapid onset of action suggested that the iVjAT-dimethyl compound itself had pharmacological effects, rather than the N-demethyl-ated metabolite, methamphetamine, although the latter is one of the known metabolites of iVjiV-dimethylamphetamine (141).
Active metabolites may be much more important in iV,iV-dialkylated compounds that possess a j3-keto function, as in cathinone (2). In that case, the AT,JV-dimethyl compound is nearly as active as the N-monomethyl compound (142). It is known, however, that the
alkyl groups of /3-aminoketones are readily cleaved metabolically. Thus, the N,N-6i-methyl cathinone analog is likely converted in vivo to the N-monomethyl compound meth-cathinone. This argument is based on evidence that for diethylpropion, the iV,2V-diethyl congener of cathinone, it is the N-monoethyl metabolite that is the active species (70, 71).
Although longer N-alkyl groups lead to less active compounds, one exception to this generalization is benzphetamine (Didrex), AT-ben-zyl-N-methylamphetamine (22). Despite the ATjAT-dialkyl groups in benzphetamine, in humans it produces subjective effects characteristic of amphetamine-like drugs such as phen-metrazine (20) (143).Althoughpara-hydroxy-N-benzylamphetamine is a major metabolite of benzphetamine, methamphetamine and amphetamine are also detectable in urine and hair following administration of benzphetamine (144-146). It is not clear from the literature whether the reinforcing effects of benz-phetamine are attributable to metabolic formation of amphetamine or methamphet-amine. Based on the studies with A^AT-dimeth-ylamphetamine by Witkin (138), however, one would predict that the parent molecule has some pharmacological activity.
The stereochemistry at the a carbon atom, when enantiomers exist, is homochiral to that of S-(+)-amphetamine (10), shown earlier. Both the releasing actions at dopamine and norepinephrine transporters in isolated rat brain slices (147) and the locomotor and stereotypic effects in rodents (148) are more potently affected by the S-(+) isomer of amphetamine than by thei?-(-) isomer. In this latter study, the (+) enantiomer was about five times more potent than the (-) isomer, paral leling the potency difference found with the enantiomers in vitro, using rat brain striatal synaptosomes (149). The two isomers were of nearly equal potency in their effects on norepi-nephrine accumulation by rat hippocampal synaptosomes (149). This stereochemical requirement applies to p-keto derivatives as well; the corresponding active isomer has the S-(-) configuration (136).
5.1.4 The a-alkyl Substituent. The a-alkyl group cannot be much larger than a methyl. Phenethylamine itself, lacking the side chain a-methyl group, is inactive in vivo because of its rapid inactivation by monoamine oxidase. Addition of the a-methyl group retards metabolism by this route, leading to the orally bio-available drug amphetamine. The uptake transporter, however, cannot tolerate large groups in this region and the a-ethyl analogs of both amphetamine and methamphetamine had markedly attenuated activity in a drug discrimination assay with rats trained to discriminate (+)-amphetamine (150). a,a-Di-methyl groups, as in phentermine (19), though giving an active compound, still reduce activity.
Attempts to incorporate the side chain into ring structures also led to compounds with attenuated activity. For example, in drug discrimination assays using rats trained to recognize the effect of (+)-amphetamine (10), compounds (23) and (24) either failed to produce amphetamine like effects, or had much lower potency (150, 151). When n = 3, the compound lacked any amphetamine-like action.
5.1.5 Other Side-Chain Substitutions. Limited substitution of the side chain is tolerated. A j3-hydroxy group on methamphetamine
gives ephedrine (1), shown earlier. Although ephedrine is a CNS stimulant, its effects are much weaker than those of methamphet-amine. Similarly, addition of a j3-hydroxy to amphetamine gives phenylpropanolamine, a compound that is nearly devoid of CNS stimulant effects. One may speculate that the polar hydroxy group reduces the hydrophobicity of these compounds such that CNS penetration is much reduced. The N-methyl of ephedrine increases lipid solubility, so ephedrine has a greater CNS action than that of phenylpropanolamine. Addition of a keto function to the structure of amphetamine or methamphet-amine gives cathinone (2) or its corresponding N-methyl derivative, methcathinone, the latter cf which also has greater potency than that cf the primary amine (142).It should be noted that an oxygen at the j3 position can be incorporated into a heterocyclic ring as in phen-metrazine (20) and phendimetrazine (21). Methyl aminorex (25) is also a potent stimulant that incorporates the essential features of the amphetamine template into an oxazoline ring. The 4S,5S-trans isomer shown (25) is the most potent of the four possible stereoisomers (152,153).
5.1.6 Aromatic Ring Substitution. Simple ring substituents can change the targets of the amphetamines from one monoamine uptake carrier to another. The dopamine and norepi-nephrine uptake carrier proteins have the most stringent structural demands, and any substitution decreases their potency at these sites. The serotonin carrier is relatively promiscuous and tolerates a variety of ring substituents, many of which dramatically increase the potency at the serotonin carrier from that of amphetamine itself. No ring modifications are known that give rise to a substituted amphetamine that completely retains amphetamine-like psychostimulant activity. para- Fluoroamphetamine (26; X=F) has been reported to have effects in rats resembling those of amphetamine, but substitution with larger halogens (e.g., chloro or iodo) leads to compounds that have significant serotonin releasing potency, and that produce behavioral effects that are different from those of amphetamine itself (154).
The i?,ii-(+)-stereoisomer of methylphenidate (8) is known to be the more active (155) and is often referred to as the active "threo" isomer. The (-)-enantiomer and the erythro stereoisomers are much less potent. One study has reported a series of aromatic ring-substituted analogs. The most potent compounds in that report were halogen substituted in the 3-or 3,4- positions of the ring. For example, the dichloro compound (27) was 32-fold more potent than methylphenidate itself in inhibiting dopamine reuptake (156). That finding parallels a recent report by Deutsch et al. (157), that replacing the phenyl ring with a j3-naph-thyl moiety (158) gave a compound with about eightfold higher affinity for the dopamine transporter. Those workers also reported that the corresponding a-naphthyl analog had only about one-tenth the potency of methylpheni-date at the DAT. Taken together, these latter observations indicate that the DAT must have a hydrophobic region that generally extends from the 3,4- positions of the aromatic phenyl ring of methylphenidate.
Deutsch et al. (157) also examined the effect of heterocyclic ring size. The pyrrolidyl
and azepino, as well as the azacyclooctane congeners, were significantly less potent than methylphenidate itself. That report also contained data for the morpholine analog of methylphenidate (158), which had an approximately 15 times lower affinity at the DAT. Beyond the studies cited here, very little additional SAR work has been done with methyl-phenidate.
Of all the psychostimulants, cocaine has probably been most studied, particularly within the last decade, as a result of its widespread abuse. Structure-activity studies have been carried out with numerous analogs, not only to elucidate the molecular requirements for interaction with the various monoamine transporters, but also in attempts to develop treatments that might be useful for cocaine addiction. Ideally, understanding the structure-activity relationships will be useful to understanding the functional topography of the binding site of the transporters, and, if a three-dimensional structure of the transporters can be developed, these features would map onto the binding site. Nevertheless, because the topic of this chapter is stimulants, and not the structure-activity relationships of monoamine transporters, an exhaustive summary of the more than 200 papers that have appeared on the SAR of cocaine and its analogs will not be presented. A useful perspective on the SAR of cocaine analogs as it was understood in 1992 has been presented by Carroll et al. (159), with more a recent update in 1997 (160).
An attempt will be made here to distill down the essence of the SAR of cocaine as it relates to its stimulant properties. In many cases, compounds have been reported that have not been tested in vivo, but have only been compared for affinity at the monoamine transporters or in an in vivo assay. Some of these data will be summarized if they are reported in the context of the stimulant effects of cocaine. Similarly, there have been numerous attempts to develop cocaine analogs that may bind to the dopamine transporter and actually block the stimulant or reinforcing effects of cocaine itself, in efforts to develop treatments for cocaine addiction. This chapter largely ignores many of those studies unless they contain in vivo data suggesting they are relevant to a discussion of stimulant effects. Nevertheless, because stimulant properties have been associated with binding to the DAT, a good deal of the SAR discussion here must be discussed in the context of in vitro DAT affinity.
A consideration of the structure-activity relationships of cocaine can focus on a number of key elements in the structure, as indicated below. Each of the following sections includes a discussion of the particular numbered structural element.
5.3.1 N-substituents. N-demethylation of cocaine has only a minor effect on affinity at monoamine transporters (161). In phenyl-tropane analogs where the ester linkage has been removed (28), extensions of the N-alkyl out to n-butyl have no effect on dopamine transporter affinity (162). Effects at the serotonin transporter are variable, but affinity only decreases modestly. At the norepinephrine transporter, affinity drops about three times with the longer N-alkyl group.
5.3.2 Basic Nitrogen Atom. For many years it was assumed that the basic nitrogen of cocaine was required for activity. It seemed logical to believe that the nitrogen, protonated at physiological pH, would interact with an an-ionic site such as an aspartate residue in the transporter (163). It was surprising, therefore, when nonbasic N-sulfonyl cocaine analogs such as (29) were found to possess high affinity for the dopamine transporter (164). These compounds are not protonated at physiological pH, and if hydrogen bonding were required for activity, these analogs could serve only as hydrogen bond acceptors. Even then, the low electron density remaining on a nitrogen with the powerfully electron-withdrawing trifluoromethylsulfonyl group attached,
would suggest that this interaction should be very weak.
Replacement of the nitrogen atom with oxygen as in (30) gives compounds that retain high affinity for the dopamine transporter (165).This finding was accommodated by proposing that the oxygen atom could act as a hydrogen bond acceptor at the transporter (165), a conclusion that would at least be consistent with the activity of the N-sulfonated derivatives (29).
It was even more surprising, therefore, when the report appeared that even a polar oxygen was not required for good uptake inhibitors. Carbocyclic compounds such as (31) proved to have transporter affinities nearly equal to those of their amine-containing counterparts (166)!
These authors postulated that there are various acceptor sites in the dopamine transporter, where an inhibitor may bind and cause uptake inhibition. The topography of these sites is probably different in the three monoamine transporters.
5,3.3 Substituent at C(2). Epimerization of the ester function to give pseudococaine (32)
results in about a 150-fold loss in affinity for the dopamine transporter (161). In compounds lacking the ester linkage (see section below) the effect is more dramatic, resulting in a more than 1000 times lower potency.
The ester is not an essential function. Replacement of the ester with an ethyl or vinyl group did not lead to significant loss of binding affinity, demonstrating that a polar function capable of hydrogen bonding was not essential (167,168). Indeed, substitution at the 2/3 position with alkyl groups as long as n-butyl, 2-phenethyl, or 2-stryl gave compounds (e.g., 33) with exceptionally high affinities at the dopamine transporter (168).
Kelkar et al. (169) have extended the 2P-alkyl group to include a polar hydroxy or methyl ester function at the distal end of a three carbon chain, with no significant loss of affinity compared to that of a simple carbome-thoxy function. They concluded that this region of the cocaine binding site must be either a large cleft in the transporter protein or exterior to the binding site. They also noted that this region is relatively insensitive to electrostatic interactions. Chang et al. (170), found that the 2P-phenyl analog (34)was equipotent to the 2-carbomethoxy compound, but had enhanced selectivity for the dopamine trans -porter over the serotonin transporter. These authors also concluded that a hydrophobic group at this region of the molecule might be a contributing factor for binding at the dopa-mine transporter.
Esters larger than a methyl are quite potent. In the 3-benzoyl series of tropane esters, both the isopropyl and phenyl esters had high affinity and selectivity for the dopamine transporter (171). The phenyl ester (35; RTI-15) dose-dependently substituted in the drug dis-
crimination paradigm in rats trained to discriminate the effects of cocaine. In contrast, whereas cocaine increased locomotor activity in mice, RTI-15 had no effect on activity and at high doses even decreased this measure (172). Because this compound was a potent inhibitor of the dopamine transporter, it suggests that high selectivity for the dopamine transporter may lead to differential retention of cocainelike effects.
The isopropyl and phenyl esters in the 3-phenyltropane series of analogs have higher affinity for the dopamine transporter than does the methyl ester (173).Similarly, tertiary amide analogs of cocaine and phenyltropane analogs are more potent than secondary or primary amides, and also have enhanced selec-
tivity for binding at the dopamine transporter over that of the norepinephrine or serotonin transporters (173). Replacement of the ester or amide function with a carboethoxy isox-azole substituted substituent gave (36), a highly potent inhibitor with selectivity for the dopamine transporter (174, 175). This compound had about twice the affinity of cocaine at the DAT.
Similarly, Carroll et al. (176) reported that the 1,2,4-oxadiazoles (e.g., 37) that are bio-isosteres of ester groups, are potent cocaine analogs. Compound (37) had low nanomolar affinity for the dopamine uptake transporter with greater than 100-fold selectivity for the transporter over the norepineph-rine and serotonin transporters.
5.3.4 The Ester Linkage at C(3). In cocaine, the 3a epimer "allococaine" (38)has considerably reduced activity compared to that of cocaine itself (177).This structural change, however, causes the tropane ring to favor the
HsC^n t\ co2ch3
pseudochair, rather than the boat, conformation that occurs in natural (-)-3j3-cocaine.
It was first reported by Clarke et al. (178) that removal of the ester linkage from cocaine, to give a compound with the phenyl ring directly attached to the tropane ring (WIN 35,065-2; 39), possessed higher affinity for the dopamine transporter than did cocaine itself. By contrast to benzoyl esters, however, the configuration at the 3 position is not so critical in phenyltropane compounds. That is, in the WIN series where the ester has been removed, the 3/3 phenyl orientation (39) was only about twofold more potent than the 3a phenyl (40) at the dopamine transporter. At the serotonin transporter, however, the 3/3 compound was significantly more potent (165). A similar trend was observed in the 8-oxa analogs, leading to the conclusion that the dopamine transporter is able to accommodate the 3-phenyl ring when the bicyclic ring is in either the boat or chair conformation, whereas the serotonin transporter is less accommodating.
5.3.5 Substitutions on the Aromatic Ring at Position 3. In the phenyltropane analogs of cocaine, where the ester linkage has been removed and the phenyl ring is attached directly to the tropane ring (WIN and RTI compounds), substitution attheparo ring position with halogens or a methyl group gave compounds (41) with increased affinities at the
dopamine transporter compared with the un-substituted compound, and with much increased affinity compared to that of cocaine itself (179). Behavioral potency paralleled the affinity increases, with all of the phenyltro-panes being considerably more potent in elevating locomotor activity in mice (180) and in substituting for a cocaine stimulus in the drug discrimination paradigm in rats (181).
The rank order of affinity for aromatic ring substituents in the WIN series was 3,4-Cl2 > I > Cl > F > H, whereas in the 8-oxa (3/3) analogs it was 3,4-Cl2 > Br > Cl > I > F > H (165). Replacing the 3j3-phenyl with either a 1- or 2-naphthyl substituent gave significantly enhanced affinity at all three monoamine transporters, with the 2-naphthyl (43) being about five- to sixfold more potent than the 1-naph-thyl (42) (182).This result is parallel to similar findings reported by Deutsch et al. (157), where replacing the phenyl ring of methyl-phenidate with a 2-naphthyl moiety gave an analog with about 70-fold higher potency than when the phenyl ring was replaced with a 1-naphthyl.
5.3.6 Requirement for the Intact Tropane Ring System. We have seen earlier that there is no absolute requirement for the basic nitrogen in the tropane structure, and that even a polar oxygen isostere replacement is not needed for cocaine congeners to possess potent monoamine reuptake inhibition. It is perhaps not too surprising, therefore, that the bridged bicyclic tropane ring is not an essential structural feature. In a series of 4-arylpiperidine carboxylic acid methyl esters, several of the compounds were significantly better uptake inhibitors than cocaine (183).The most potent compound in the series, (44), was about 20 times more potent at the dopamine uptake transporter than cocaine.
Was this article helpful?